1. Volume 16, Issue 5, Pages 504-516 (May 2015)
Opposing Roles for the lncRNA Haunt and Its Genomic Locus in Regulating HOXA Gene Activation during Embryonic Stem Cell Differentiation 
Yafei Yin, Pixi Yan, Jinlong Lu, Guang Song, Yangyang Zhu, Zhaohui Li, Yi Zhao, Bin Shen, Xingxu Huang, Heng Zhu, Stuart H. Orkin, Xiaohua Shen 
Cell Stem Cell 
Volume 16, Issue 5, Pages (May 2015)
DOI: /j.stem
Copyright © 2015 Elsevier Inc. Terms and Conditions

2. Cell Stem Cell 2015 16, 504-516DOI: (10.1016/j.stem.2015.03.007)
Copyright © 2015 Elsevier Inc. Terms and Conditions

3. Figure 1 CRISPR-Mediated Knockout of the Haunt Locus
(A) Schematic diagram of ESC differentiation and RNA-seq analysis.
(B) Heat map of FPKM values of lncRNAs enriched in ESCs and lineage-committed cells. Here, we show 83 genes enriched in ESCs (D0); 130, 134, and 65 genes at days 2, 4, and 6 of ESC differentiation induced by LIF withdrawal (D2, D4, and D6, respectively); 65 in mesendoderm (ME) cells; 72 in neural progenitor cells (NPCs); and 122 in neural stem cells (NSCs). See also Table S1A.
(C) Northern blot analysis showing a major Haunt transcript.
(D) RT-qPCR analysis of Haunt RNAi. CTL represents the scrambled shRNA control. n = 4, including 2 shRNAs and 2 technical replicates. ∗∗p < 0.01.
(E) Schematic diagram of knockout strategies at the Haunt and HOXA loci. Haunt exons (E1–E3) are shown in black boxes. A targeting vector for Haunt 58kb KO is shown at the bottom. P1 and P2 are Southern blot probes.
(F) Validation of knockout ESC lines by PCR and Southern blotting analysis. Gel images from representative clones (a–m) are shown. CTL, an unrelated genomic amplicon; 5′ and 3′ HA, PCR to detect homologous recombination at the 5′ and 3′ ends of the homology arms; WT, wild-type allele; KO, knockout allele.
(G) RT-qPCR analysis of Haunt in knockouts. “q1,” primers q1f/q1r; “q2,” q2f/q2r. n = 2 replicates.
In (D) and (G), data are shown as mean ± SD. See also Figures S1 and S2 and Table S1.
Cell Stem Cell  , DOI: ( /j.stem )
Copyright © 2015 Elsevier Inc. Terms and Conditions

4. Figure 2 Misregulation of HOXA Genes in Haunt Loss-of-Function Mutants
(A) Time-course RT-qPCR analysis of RA treatment in wild-type (WT) and Haunt 58kb KO ESCs. n = 4, including 2 independent clones and 2 technical replicates.
(B and C) RT-qPCR analysis of HOXA genes in Haunt knockdown (B) and knockout (C) ESCs at day 1 of RA treatment.
(D) Heat maps of FPKM values of Haunt and HOX genes in day 2 and day 4 RA-treated ESCs. Genes with FPKM > 2 in at least one sample are shown. See also Table S2A.
(E) Schematic diagram of Haunt 4×pA knockin.
(F) RT-qPCR analysis of HOXA genes in Haunt 4×pA KI cells.
In (A)–(C) and (F), data are shown as mean ± SD. n = 4 or 6, including 2 shRNAs for knockdown, 3 ESC clones (#a, c, and d) for pKO, 2 independent clones for other knockouts, 3 independent clones for 4×pA KI, and 2 technical replicates per ESC clone. ∗p < 0.05 and ∗∗p < See also Figure S3 and Table S2.
Cell Stem Cell  , DOI: ( /j.stem )
Copyright © 2015 Elsevier Inc. Terms and Conditions

5. Figure 3 CRISPR-Mediated Knockin Analysis of Haunt Function
(A) Schematic diagram of CAG KI.
(B) Southern blotting analysis of Haunt CAG KI cells.
(C) RT-qPCR analysis in Haunt CAG KI ESCs.
(D) RT-qPCR analysis of HOXA genes in Haunt CAG KI cells at day 1 of RA treatment.
(E) Schematic diagram of Haunt P-Inr2 KO and P-Inr2/CAG-cDNA KI cells. A region spanning from the promoter to the end of intron 2 (P-Inr2) was targeted to simultaneously establish deletion and knockin alleles.
(F) Southern blotting analysis of representative Haunt P-In2/CAG-cDNA KI and P-Inr2 KO ESCs. WT, wild-type allele; CAG KI, CAG-cDNA knockin allele; KO, allele with a deletion occurring at or near the expected Cas9 cutting sites. Clone “a” of the P-Inr2 KO ESCs contained an unknown mutation and was thus excluded from the subsequent analysis.
(G) RT-qPCR analysis of Haunt and HOXA genes at day 1 of RA treatment.
(H) Schematic summary of the effect of genetic manipulation of the Haunt locus on HOXA induction. The upper panel shows the relative location of the Haunt and HOXA genes in the genome. Haunt pKO cells exhibit enhanced activation, whereas all other cells exhibit downregulated expression of HOXA genes. Brown bars with double arrowheads indicate relative deletion regions. Green arrows indicate the knockin of a CAG promoter or a CAG-cDNA. Means of fold changes relative to the expression in wild-type ESCs were plotted.
In (C), (D), and (G), data are shown as mean ± SD. (n = 4, including 2 independent clones and 2 technical replicates.) ∗p < 0.05 and ∗∗p < See also Figure S3.
Cell Stem Cell  , DOI: ( /j.stem )
Copyright © 2015 Elsevier Inc. Terms and Conditions

6. Figure 4
Haunt DNA Contains Enhancers and Interacts with the HOXA Region on Chromatin
(A) The Haunt DNA locus contains putative enhancer sequences. Deletion regions are shown at the top. ∗ indicates anchor primers, and the black arrow shows the direction of 3C primers located upstream of each selected HindIII cutting site. All sequencing tracks are from undifferentiated (D0) or RA-treated ESCs, except for H3K27ac ChIP-seq in RA-treated P19 embryonic carcinoma cells. RARa-binding peaks are shown in vertical bars at the bottom.
(B and C) 3C analysis of chromatin interactions between the Haunt and HOXA loci. The y axis shows the interaction frequency normalized to the control BAC DNA and the nearest upstream site, #27 in (B), #52 in (C). Data are shown as mean ± SEM (n = 3, experimental replicates). The p values between indicated samples in the region #6–14 (green shadow) are shown. Two additional 3C experiments showed similar results (Figure S5).
See also Figure S5.
Cell Stem Cell  , DOI: ( /j.stem )
Copyright © 2015 Elsevier Inc. Terms and Conditions

7. Figure 5 Haunt RNA Binds to HOXA Loci on Chromatin
(A) Subcellular fractionation of Haunt. Gapdh, U1, and Xist RNAs serve as controls for the cytosolic, nuclear, and chromatin fractions, respectively.
(B) Scheme illustrating chromatin isolation by RNA purification (ChIRP) of Haunt.
(C) Deep sequencing tracks of Haunt ChIRP, H3K27me3, and H3K4me3 ChIP in undifferentiated (D0) and day 1 RA-treated (RA) ESCs. The upper tracks show normalized read densities of RNA-seq and DNA-seq following the ChIRP procedure. WT, wild-type ESCs; KO, P-exon2 KO as the control for ChIRP (with the deletion region shown at the top). The RNA-seq track of RA-treated ESCs on day 1 is shown at the bottom.
(D) A zoomed-out view of Haunt ChIRP DNA-seq signals in wild-type ESCs within a 1-Mb range of the Haunt locus. The Haunt TSS is centered in the middle (the red arrow). Normalization was performed by dividing the number of actual reads in each 10-kb bin by the sequencing depth. The ChIRP DNA-seq tracks are shown at the bottom.
(E) Quantification of Haunt ChIRP DNA-seq signals in the genomic regions of Haunt and nearby genes in wild-type ESCs. The fold change between D0 and RA-treated ESCs is shown above the corresponding gene.
(F) Quantification of H3K27me3 ChIP-seq signals in regions of individual genes. Brackets indicate genes with >1.8-fold change between RA-treated wild-type and pKO cells.
(G) H3K27me3 ChIP-qPCR. The y axis shows fold enrichment relative to the input and genomic background (g-CS). Data are shown as mean ± SD. (n = 3 biological replicates.)
In (E) and (F), the y axis shows normalized numbers of reads that fall into the DNA locus of each gene, covering the 1-kb region upstream and downstream of the gene body, to the sequencing depth and the length of each gene region. See also Figures S6 and S7.
Cell Stem Cell  , DOI: ( /j.stem )
Copyright © 2015 Elsevier Inc. Terms and Conditions

8. Figure 6 Haunt Orchestrates Proper ESC Differentiation
(A) Gene set enrichment analysis (GSEA) profiles on day 4 of RA treatment. Compared with the wild-type ESCs, Haunt 58kb KO, CAG KI, and HOXA KO, but not pKO, ESCs showed significant depletion or downregulation of RA-induced genes. NES, normalized enrichment scores. See also Table S3.
(B) Gene ontology (GO) analysis of genes that were downregulated in HOXA KO and Haunt 58kb KO ESCs compared with the wild-type on day 4 of RA treatment. See also Table S4.
(C) Heat map of FPKM values of representative genes that are misregulated in various mutant ESCs on day 2 or day 4 of RA treatment. See also Table S2B.
(D) A model showing the complex effects of Haunt RNA/transcription and its DNA locus in fine-tuning HOXA expression and orchestrating ESC differentiation. The Haunt DNA locus provides enhancers that are required for the activation of HOXA genes, whereas Haunt RNA transcripts attenuate long-range chromatin interactions between the Haunt enhancer DNA and the HOXA region likely through its direct chromatin association. In addition, Haunt RNA and/or transcription may reshape the H3K27me3 landscape at the HOXA loci during RA-induced ESC differentiation. Thus, in opposing the activating role of Haunt DNA sequences, Haunt RNAs serve as a “brake” to prevent aberrant activation of HOXA genes. The fine balance between the active and repressive functions of Haunt DNA and RNA, respectively, precisely controls the proper expression of the developmentally regulated HOXA locus and contributes to orchestrated differentiation of ESCs. See also Tables S2, S3, and S4.
Cell Stem Cell  , DOI: ( /j.stem )
Copyright © 2015 Elsevier Inc. Terms and Conditions